1:
2: \documentclass{elsart}
3: \usepackage{natbib}
4: \usepackage{amsmath}
5: \usepackage{amssymb}
6: \usepackage{graphicx}
7:
8: \begin{document}
9:
10: \begin{frontmatter}
11:
12:
13:
14: \title{Exponentially Localized Solutions of Mel'nikov
15: Equation}
16:
17: \author{C. Senthil Kumar\thanksref{label1}},
18: \author{ R. Radha\thanksref{label2}},
19: \author{M. Lakshmanan
20: \thanksref{label1}\corauthref{cor1}}
21: \ead{lakshman@cnld.bdu.ac.in}
22: \corauth[cor1]{Corresponding Author. Fax:+91-431-2407093}
23: \address[label1]{Centre for Nonlinear Dynamics, Dept. of Physics, Bharathidasan University,
24: Tiruchirapalli -620 024, India.}
25: \address[label2]{Dept. of Physics, Govt. College for Women, Kumbakonam - 612 001,
26: India.}
27:
28: \begin{abstract}
29: The Mel'nikov equation is a (2+1) dimensional nonlinear evolution equation
30: admitting boomeron type solutions. In this paper, after showing that it
31: satisfies the Painlev\'{e} property, we obtain exponentially localized dromion
32: type solutions from the bilinearized version which have not been reported so far.
33: We also obtain more general dromion type solutions with spatially varying amplitude as well as
34: induced multi-dromion solutions.
35: \end{abstract}
36:
37:
38: \end{frontmatter}
39:
40:
41: \section{Introduction}
42: The identification of dromions which are exponentially localized solutions
43: in (2+1) dimensional soliton equations [1-6] has been one of the most
44: interesting developments in soliton theory in recent times, which has given
45: a fillip to the understanding of integrable systems in (2+1) dimensions.
46: Essentially, these localized solutions arise due to the presence of some
47: additional nonlocal terms or effective local fields associated with
48: ``boundaries''. Further, the advent of ``explode decay dromions'' which are
49: again exponentially localized solutions with time varying amplitudes [7]
50: and induced dromions [6,8] using arbitrary functions of space and time
51: variables have set in motion the process of unearthing more and more
52: novel localized entities in (2+1) dimensional nonlinear systems.
53:
54:
55: \subsection{Mel'nikov Equation}
56: An interesting evolution equation in (2+1) dimensions which we consider here is the
57: one proposed by Mel'nikov [9,10] that describes (under certain conditions)
58: the interaction of two waves on the x-axis. This equation is of the form
59: \begin{subequations}
60: \begin{eqnarray}
61: 3u_{tt}-[u_{y}+(3u^2+u_{xx}+8\kappa|\chi|^2)_{x}]_{x} = 0, \\
62: i\chi_{t} = u\chi+\chi_{xx},
63: \end{eqnarray}
64: \end{subequations}
65:
66:
67:
68: where $u$ is the long wave amplitude (real), $\chi$ is the complex short
69: wave envelope, and the parameter $\kappa$ satisfies the condition $\kappa^2$ =
70: 1. Eq. (1) may be considered either as a generalization of the K-P equation
71: with the addition of a complex scalar field or as a generalization of the NLS
72: equation with a real scalar field (after suitable interchange of coordinates $y$
73: and $t$). Mel'nikov [10] has pointed out that Eq. (1)
74: admits boomeron type solutions, which can be realized from an asymptotic
75: analysis of the two soliton solution.
76:
77: It is expected that the investigation of this equation may have wider
78: ramifications in plasma physics, nonlinear optics and hydrodynamics. It is
79: this diverse presence of this equation which prompts one to make a detailed
80: investigation of their dynamics, particularly to identify whether localized
81: solutions exist in this system. For this purpose, we first carry out a
82: Painlev\'e singularity structure analysis and confirm that Eq. (1) does indeed
83: satisfy the Painlev\'e property. Then bilinearizing the evolution equation and
84: making use of certain arbitrary functions present in the solution, we obtain a
85: large class of exponentially localized dromion solutions.
86:
87: \section{Singularity Structure Analysis of Mel'nikov Equation}
88: We explore the singularity structure of Eq. (1), by rewriting
89: $\chi$=a and $\chi^*$=b as
90: \begin{subequations}
91: \begin{eqnarray}
92: 3u_{tt}-u_{xy}-6u_{x}^2-6uu_{xx}-u_{xxxx}-8\kappa(a_{xx}b+2a_{x}b_{x}+ab_{xx})=0,\\
93: ia_{t}=ua+a_{xx},\\
94: -ib_{t}=ub+b_{xx}.
95: \end{eqnarray}
96: \end{subequations}
97: We now effect a local Laurent expansion in the neighbourhood of a
98: noncharacteristic singular manifold $\phi(x,y,t)$ = 0,
99: $\phi_{x}\neq$ 0, $\phi_{y}\neq$ 0. Assuming the leading orders of the solutions
100: of Eq. (2) to have the form
101: \begin{equation}
102: u=u_{0}\phi^{\alpha},a=a_{0}\phi^{\beta},b=b_{0}\phi^{\gamma},
103: \end{equation}
104: where $u_{0}$, $a_{0}$ and $b_{0}$ are analytic functions of ($x$, $y$, $t$) and
105: $\alpha$, $\beta$, $\gamma$ are integers to be determined, we now substitute
106: (3) into (2) and balance the most dominant terms to get
107: \begin{equation}
108: \alpha=\beta=\gamma=-2,
109: \end{equation}
110: with the condition
111: \begin{equation}
112: \kappa a_{0}b_{0}=-9\phi_{x}^4, u_{0}=-6\phi_{x}^2.
113: \end{equation}
114: Now, considering the generalized Laurent expansion of the solutions in the
115: neighbourhood of the singular manifold
116: \begin{subequations}
117: \begin{eqnarray}
118: u=u_{0}\phi^{\alpha}+...+u_{r}\phi^{r+\alpha}+..., \\
119: a=a_{0}\phi^{\beta}+...+a_{r}\phi^{r+\beta}+..., \\
120: b=b_{0}\phi^{\gamma}+...+b_{r}\phi^{r+\gamma}+...,
121: \end{eqnarray}
122: \end{subequations}
123: the resonances (powers) at which arbitrary functions enter into (6) can be
124: determined by substituting (6) into (2) and comparing the coefficients
125: of ($\phi^{r-6}$,$\phi^{r-4}$,$\phi^{r-4}$) to give
126: \begin{eqnarray}
127: \left(
128: \begin{array}{ccc}
129: \phi_{x}^2(r^2-5r-30)\hat{r} & 8\kappa b_{0}\hat{r} &
130: 8\kappa a_{0}\hat{r} \\
131: a_{0} & \phi_{x}^2(r^2-5r) & 0 \\
132: b_{0} & 0 & \phi_{x}^2(r^2-5r)
133: \end{array}
134: \right)
135: \left(
136: \begin{array}{c}
137: u_{r} \\
138: a_{r} \\
139: b_{r}
140: \end{array}
141: \right) = 0,
142: \end{eqnarray}
143: where $\hat{r}=(r-4)(r-5)$. Solving Eq. (7), one gets the resonance values as
144: \begin{equation}
145: r=-3,-1,0,4,5,5,6,8.
146: \end{equation}
147:
148: The resonance at $r$ = -1 naturally represents the arbitrariness of the manifold
149: $\phi(x,y,t)=0$. In order to prove the existence of arbitrary functions at the other resonance
150: values, we now substitute the full Laurent series
151: \begin{subequations}
152: \begin{eqnarray}
153: u=u_{0}\phi^{\alpha}+\sum_{r}u_{r}\phi^{r+\alpha}, \\
154: a=a_{0}\phi^{\beta}+\sum_{r}a_{r}\phi^{r+\beta}, \\
155: b=b_{0}\phi^{\gamma}+\sum_{r}b_{r}\phi^{r+\gamma}
156: \end{eqnarray}
157: \end{subequations}
158: into Eq. (2). Now collecting the coefficients of ($\phi^{-6}$,$\phi^{-4}$,$\phi^{-4}$) and
159: solving them, we obtain the relations (5), implying a resonance at $r$ = 0.
160:
161: Similarly collecting the following coefficients, we obtain the necessary
162: information about the positive resonances:
163:
164: (i) coefficients of ($\phi^{-5}$,$\phi^{-3}$,$\phi^{-3}$):
165: \begin{eqnarray}
166: u_{1}=0,
167: a_{1}=\frac{ia_{0}\phi_{t}}{2}-a_{0x},
168: b_{1}=\frac{-ib_{0}\phi_{t}}{2}-b_{0x}.
169: \end{eqnarray}
170:
171: (ii) coefficients of ($\phi^{-4}$,$\phi^{-2}$,$\phi^{-2}$): $u_{2}$, $a_{2}$,
172: and $b_{2}$ can be uniquely determined.
173:
174: (iii) coefficients of ($\phi^{-3}$,$\phi^{-1}$,$\phi^{-1}$): $u_{3}$, $a_{3}$,
175: and $b_{3}$ can be uniquely determined.
176:
177: (iv) coefficients of ($\phi^{-2}$,$\phi^{0}$,$\phi^{0}$): Only two equations
178: result for three unknowns $u_{4}$, $a_{4}$, $b_{4}$ and so one of them is
179: arbitrary, corresponding to a resonance at $r$ = 4.
180:
181: (v) coefficients of ($\phi^{-1}$,$\phi^{1}$,$\phi^{1}$): Only $u_{5}$ is determined,
182: while $a_{5}$ and $b_{5}$ are arbitrary corresponding to double resonance
183: at $r$ = (5, 5).
184:
185: (vi) coefficients of ($\phi^{0}$,$\phi^{2}$,$\phi^{2}$): Only two equations
186: result for three unknowns $u_{6}$, $a_{6}$, $b_{6}$ and so one of them is
187: arbitrary, corresponding to $r$ = 6.
188:
189: (vii) coefficients of ($\phi^{1}$,$\phi^{3}$,$\phi^{3}$): $u_{7}$, $a_{7}$,
190: and $b_{7}$ can be determined in terms of earlier coefficients.
191:
192: (viii) coefficients of ($\phi^{2}$,$\phi^{4}$,$\phi^{4}$): Only two equations
193: result for three unknowns $u_{8}$, $a_{8}$, $b_{8}$ and so one of them is
194: arbitrary, corresponding to $r$ = 8.
195:
196: For the negative resonance $r$ = -3, following the approach of Conte, Fordy and
197: Pickering [11], we demand that both the solution of Mel'nikov Eq. (2) and
198: the solution close to it represented by a peturbation series in a small parameter
199: $\epsilon$ are free from movable critical manifolds. We identify that the first
200: order peturbed series does admit an
201: arbitrary function corresponding to a movable pole at the resonance $r$ = -3.
202: Consequently, for each of the eight resonances given by Eq. (8), one can
203: associate an arbitrary function in the solution (and close to it) without the
204: introduction of movable critical manifolds.
205:
206: It must be mentioned that the above system (2) admits another leading order
207: behaviour with $\alpha=-2$, $\beta=\gamma=-1$, $u_{0}=-2\phi_{x}^2$, $a_{0}$
208: and $b_{0}$ are arbitrary. The Laurent series with the above leading order
209: leads to resonances $0,0,-1,3,3,4,5,6$, corresponding to a normal branch
210: and the existence of sufficient number
211: of arbitrary functions can be established at these resonance values. This
212: can also be verified from the corresponding bilinear form as studied by
213: Grammaticos, Ramani and Hietarinta [12].
214: Consequently one can be assured that the Mel'nikov equation (1) or (2)
215: indeed satisfies the Painlev\'e property.
216:
217: \section{Bilinearization of Mel'nikov Equation and Localized Solutions}
218: We next Hitora bilinearize the Mel'nikov Eq. (1) to bring out the existence
219: of exponentially localized solutions of Mel'nikov equation.
220: Making the transformation
221: \begin{subequations}
222: \begin{eqnarray}
223: u &=& 2\frac{\partial^2}{\partial x^2}\mbox{ln}F,\\
224: \chi &=& \frac{G}{F},
225: \end{eqnarray}
226: \end{subequations}
227: identifiable from the Painlev\'e analysis, Eq. (1) gets converted into the
228: following Hirota bilinear form,
229: \begin{subequations}
230: \begin{eqnarray}
231: (3D_{t}^2-D_{x}D_{y}-D_{x}^4)F.F = 8\kappa|G|^2, \\
232: iD_{t}G.F = D_{x}^2G.F,
233: \end{eqnarray}
234: \end{subequations}
235: where the $D$'s are the usual bilinear operators. However, this bilinearization
236: have been already done by Y. Hase etal [13] where they have given soliton
237: solutions whereas we have brought out localised solutions here. For a completion,
238: we proceed as follows. Introducing the series expansion,
239: \begin{subequations}
240: \begin{eqnarray}
241: G &=& \epsilon g^{(1)}+\epsilon^3 g^{(3)}+..., \\
242: F &=& 1+\epsilon^2 f^{(2)}+\epsilon^4 f^{(4)} +...
243: \end{eqnarray}
244: \end{subequations}
245: into the above bilinear form and gathering terms with various powers of the
246: small parameter $\epsilon$, we obtain the following set of equations,
247: \begin{subequations}
248: \begin{eqnarray}
249: O(\epsilon) &:& ig_{t}^{(1)} = g_{xx}^{(1)}, \\
250: O(\epsilon^2) &:& 3 f_{tt}^{(2)}-f_{xy}^{(2)}-f_{xxxx}^{(2)} = 4\kappa
251: g^{(1)}g^{(1)*},
252: \end{eqnarray}
253: \end{subequations}
254: etc. Solving Eq. (14a), we can immediately write down the following solution,
255: \begin{equation}
256: g^{(1)} = \sum_{j=1}^{N}exp(\psi_{j}), \psi_{j} =
257: l_{j}x+m_{j}y+\omega_{j}t+\psi_{j}^{(0)}, i\omega_{j}=l_{j}^2,
258: \end{equation}
259: where the spectral parameters $l_{j}, m_{j}, \omega_{j}$ and $\psi_{j}^{(0)}$ are
260: all complex. Confining to $N$ = 1 in Eq. (15) and substituting (15) into
261: (14b), we obtain
262: \begin{equation}
263: f^{(2)} = \mbox{exp}(\psi_{1}+\psi_{1}^*+2A),\; \mbox{exp}(2A) = {\kappa\over
264: {(3\omega_{1R}^2-l_{1R}m_{1R}-4l_{1R}^4)}}.
265: \end{equation}
266: Here $l_{1}=l_{1R}+il_{1I}$, $m_{1}=m_{1R}+im_{1I}$ and $\omega_{1}=
267: \omega_{1R}+i\omega_{1I}$ and also $i\omega_{1}=l_{1}^2$.
268: Choosing $g^{(2j+1)}=0, f^{(2j)}=0$, for
269: $j$ $>$ 1, in Eq. (13) and using Eqs. (15) and (16) alongwith
270: the transformation (11), the physical field $u$ and the potential $\chi$
271: can be easily seen to be driven by the envelope soliton and pulse soliton
272: respectively as
273: \begin{subequations}
274: \begin{eqnarray}
275: \chi &=& \left(\frac{\sqrt{l_{1R}}(12l_{1R}l_{1I}^2-m_{1R}-4l_{1R}^3)^
276: {\frac{1}{2}}}{2\sqrt\kappa}\right)
277: \mbox{sech}(\psi_{1R}+A)e^{i\psi_{1I}}, \\
278: u &=& 2l_{1R}^2\mbox{sech}^2(\psi_{1R}+A).
279: \end{eqnarray}
280: \end{subequations}
281: where $\psi_{1R}=l_{1R}x+m_{1R}y+2l_{1R}l_{1I}t+\psi_{1R}^{(0)}$ and
282: $\psi_{1I}=l_{1I}x+m_{1I}y-(l_{1R}^2-l_{1I}^2)t+\psi_{1I}^{(0)}$. One can proceed further in
283: the standard way to obtain higher order soliton solutions also.
284:
285: \subsection{Dromions}
286: Looking at the above solutions (17), we realize the fact that as the parameter
287: $l_{1R}$ $\rightarrow$ 0, both $u$ and $\chi$ vanish. But, when $m_{1R}$
288: $\rightarrow$ $12l_{1R}l_{1I}^2-4l_{1R}^3$, the potential $\chi$ vanishes,
289: whereas the physical field $u$ survives and is driven by a ghost soliton
290: of the form,
291: \begin{equation}
292: u=2l_{1R}^2\mbox{sech}^2(\psi_{1R}+c),
293: \end{equation}
294: where $c$ is a new constant. This predicts the existence of
295: exponentially localized solution for the complex field variable $\chi$
296: in Eq. (1).
297:
298: \subsubsection{(1,1) Dromion}
299: To generate a (1,1) dromion, we now make the ansatz
300: \begin{equation}
301: F = 1+e^{\tilde{\psi}_{1}+\tilde{\psi}_{1}^*}+e^{\tilde{\psi}_{2}+\tilde{\psi}_{2}^*}
302: +K e^{\tilde{\psi}_{1}+\tilde{\psi}_{1}^*+\tilde{\psi}_{2}+\tilde{\psi}_{2}^*}, \\
303: \end{equation}
304: where
305: \begin{subequations}
306: \begin{eqnarray}
307: \tilde{\psi}_{1} &=& px+\omega t, \\
308: \tilde{\psi}_{2} &=& qy.
309: \end{eqnarray}
310: \end{subequations}
311: Here $K$ is a real constant and $p, \omega$ and $q$ are complex constants.
312: Substituting (19) in (12a), we obtain
313: \begin{subequations}
314: \begin{eqnarray}
315: G = \rho e^{\tilde{\psi}_{1}+\tilde{\psi}_{2}}, \\
316: 4\kappa |\rho|^2 = (p+p^*)(q+q^*)(1-K)
317: \end{eqnarray}
318: \end{subequations}
319: for the parametric choice $3\omega_{R}^2=4p_{R}^4$. Substituting (19) and (21)
320: in (12b), we find that $\omega=-ip^2$, where $p_{R}=\pm\sqrt{3}p_{I}$,
321: $p_{R}$ and $p_{I}$ are the real and imaginary parts, respectively of $p$.
322: Hence, the exponentially localized solution with one bound state for the
323: potential field $\chi$ for the above choice of $\omega$ and $p$ takes the form
324: \begin{equation}
325: \chi = {\rho e^{\tilde{\psi}_{1}+\tilde{\psi}_{2}}\over{1+e^{\tilde{\psi}_{1}
326: +{\tilde{\psi}_{1}}^*}+e^{\tilde{\psi}_{2}+{\tilde{\psi}_{2}^*}}
327: +K e^{\tilde{\psi}_{1}+{\tilde{\psi}_{1}}^*+\tilde{\psi}_{2}+{\tilde{\psi}_{2}}^*}}},\\
328: \end{equation}
329: while the scalar field has the form
330: \begin{equation}
331: u = {4p_{R}^2(1+e^{\tilde{\psi}_{2}+{\tilde{\psi}_{2}}^*})
332: (e^{\tilde{\psi}_{1}+{\tilde{\psi}_{1}}^*}+Ke^{\tilde{\psi}_{1}+{\psi_{1}}^*
333: +\tilde{\psi}_{2}
334: +{\tilde{\psi}_{2}}^*})\over{(1+e^{\tilde{\psi}_{1}+{\tilde{\psi}_{1}}^*}
335: +e^{\tilde{\psi}_{2}+{\tilde{\psi}_{2}}^*}+Ke^{\tilde{\psi}_{1}+{\tilde{\psi}_{1}}^*+\tilde{\psi}_{2}
336: +{\tilde{\psi}_{2}}^*})^2}},
337: \end{equation}
338: which is always bounded, but localized everywhere except in the neighbourhood
339: of the line $\tilde{\psi}_{1}+\tilde{\psi}_{1}^*=0$ in the $(x,y)$ plane. A snapshot of
340: \begin{figure}[!ht]
341: \begin{center}
342: \includegraphics[width=0.75\linewidth]{fig1}
343: \end{center}
344: \caption{Snapshot of the (1,1) dromion solution of the Mel'nikov equation
345: (see Eq.~(22))}
346: \end{figure}
347: the (1,1) dromion solution for the magnitude of the potential field $\chi$
348: is shown in Fig. 1. One can proceed to find multi-dromion solutions also,
349: generalizing the above (1,1) dromions.
350:
351: \subsubsection{Dromions with Spatially Varying Amplitude}
352: It can be seen from Eq. (14a) that the differential equation for $g$ involves only
353: the variables $t$ and $x$. Hence, an arbitrary function of $y$ can also enter
354: into its solution so that the most general form of it can be given as
355: \begin{equation}
356: g^{(1)} = \sum_{j=1}^{N}exp(\tilde{\psi}_{j}), \tilde{\psi}_{j} =
357: l_{j}x+f_{j}(y)+w_{j}t+\psi_{j}^{(0)}, iw_{j}=l_{j}^2,
358: \end{equation}
359: where $f_{j}(y)$s are arbitrary functions of $y$.
360: This fact can be harnessed in a suitable way to construct
361: a more general class of localized solutions. Following the above procedure
362: to derive the (1,1) dromion solution (22), one can easily
363: obtain the generalized dromion solution involving arbitrary function of
364: $y$ in the same form as Eq. (22) except that $\tilde{\psi}_{2}$ is now given by
365: \begin{equation}
366: \tilde{\psi}_{2} = f(y),
367: \end{equation}
368: instead of (20b), where the arbitrary function $f(y)$ is in general complex. The amplitude of
369: the localized solution with arbitrary function of $y$ is now defined by the
370: equation
371: \begin{equation}
372: 4\kappa |\rho|^2=2(p+p^*)(1-K)f'_{R}(y).
373: \end{equation}
374: where $f'_{R}(y)$ is the derivative of the real part of $f$ with respect to $y$.
375: Thus, the amplitude of the above localized solution varies with the spatial
376: coordinate $y$ by virtue of Eq. (26). This situation is reminiscent
377: of the explode-decay dromion of the variable coefficient DSI equation [7]
378: where the amplitude varies with time. It should be mentioned that, to our
379: knowledge this is the first time the amplitude of a localized solution of a
380: (2+1) dimensional nonlinear partial differential equation has been found to
381: vary as a function of the spatial coordinate $y$. This can be easily
382: generalized to multi-dromions with spatially varying amplitude.
383:
384:
385: \subsubsection{Induced dromions}
386: We also wish to point out that the existence of an arbitrary function in the solution
387: of $g^{(1)}$ in Eq. (14a) can be further utilized to obtain new induced
388: dromion solutions [6] for $\chi$.
389: For example, Eqs. (14a) and (14b) can also be solved in terms of
390: arbitrary functions as
391: \begin{equation}
392: g^{(1)}=a(y)e^{\hat{\psi_{1}}},\hat{\psi_{1}}=l_{1}x+\omega_{1}t,
393: \end{equation}
394: where $a(y)$ is an arbitrary complex function of $y$ and $l_{1}$ and
395: $\omega_{1}$ are complex constants constrained by the condition
396: $i\omega_{1}=l_{1}^2$. Substituting the form (27) into
397: (14b) and solving the resultant equation, we obtain
398: \begin{equation}
399: f^{(2)}=b(y)e^{\hat{\psi_{1}}+\hat{\psi_{1}}^*},
400: \end{equation}
401: where $b(y)$ is a real function of $y$ and is given by the condition
402: \begin{equation}
403: 6\omega_{1R}^2b(y)-l_{1R}b_{y}-8l_{1R}^4b(y) = 2\kappa|a(y)|^2.
404: \end{equation}
405: The above solutions can then be used to generate curved line soliton for the
406: field variable and the potential as
407: \begin{subequations}
408: \begin{eqnarray}
409: \chi &=& {a(y) \over {2\sqrt{b(y)}}}e^{i\hat{\psi_{1I}}}\mbox{sech}[\hat{\psi_{1R}}+
410: {1\over{2}}\mbox{log}b(y)],\\
411: u &=& 2l_{1R}^2\mbox{sech}^2[\hat{\psi_{1R}}+{1\over{2}}\mbox{log}b(y)].
412: \end{eqnarray}
413: \end{subequations}
414: where $\hat{\psi_{1R}}$ and $\hat{\psi_{1I}}$ are the real and imaginary parts,
415: repectively of $\hat{\psi_{1}}$. Thus, by choosing the arbitrary
416: functions $a(y)$ and $b(y)$ suitably which are constrained by the Eq. (29),
417: one can induce localized solutions for the field variable $\chi$ as in the case of
418: Zakharov-Strachan equation [6]. Eventhough there exists two functions $a(y)$
419: and $b(y)$, only one of them is found to be arbitrary, which is evident from the
420: Eq. (29) above. For example, by choosing
421: \begin{equation}
422: {a(y)\over \sqrt{b(y)}} = 2\mbox{sech}(m_{1}y),
423: \end{equation}
424: where $m_{1}$ is a real constant, we can find from Eq. (29) that
425: \begin{equation}
426: b(y)=\mbox{exp}\left(-\frac{1}{l_{1R}}(8\kappa\mbox{tanh}m_{1}y+(8l_{1R}^4
427: -6\omega_{1R}^2)y)\right).
428: \end{equation}
429: Then we obtain the induced localized solution
430: \begin{equation}
431: \chi=e^{i\hat{\psi_{1I}}}\mbox{sech}(m_{1}y)\mbox{sech}\bigg[l_{1R}(x
432: +2l_{1I}t)-\frac{1}{2l_{1R}}\bigg(8\kappa\mbox{tanh}m_{1}y+(8l_{1R}^4-6\omega_{1R}^2)y
433: \bigg)\bigg].
434: \end{equation}
435:
436: Thus, by choosing $a(y)$ and $b(y)$ suitably, one can induce a wide class of
437: localized solutions for the Mel'nikov Eq. (1).
438: For example, choosing an algebraic form
439: \begin{equation}
440: {a(y)\over \sqrt{b(y)}}={2\over{(y+y_{0})^2+1}},
441: \end{equation}
442: we obtain an algebraically decaying localized solution
443: \begin{eqnarray}
444: \chi &=&{1\over{(y+y_{0})^2+1}}e^{i\hat{\psi_{1I}}}\mbox{sech}\bigg[l_{1R}(x
445: +2l_{1I}t)-\frac{1}{2l_{1R}}\bigg((8l_{1R}^4-6\omega_{1R}^2)y \nonumber\\
446: & &
447: +8\kappa\int
448: \frac{dy}{((y+y_{0})^2+1)^2}\bigg)\bigg].
449: \end{eqnarray}
450: One can as well generalize this procedure to construct even wider class of
451: localized solutions. In fact, multi-induced dromions take the simple form
452: \begin{equation}
453: \chi_{N}=\left[{\sum_{j=1}^{N}a_{j}
454: (y)\over 2\sqrt{b(y)}}\right]
455: e^{i\hat{\psi_{1I}}}\mbox{sech}[\hat{\psi_{1R}}+{1\over{2}}\mbox{log}b(y)],
456: \end{equation}
457: where $a_{j}$, $j=1,2...N$ are arbitrary functions of $y$ and they are related
458: to $b(y)$ by the relation
459: \begin{equation}
460: 6\omega_{1R}^2b(y)-l_{1R}b_{y}-8l_{1R}^4b(y) = 2\kappa\sum_{p=1}^{N}
461: \sum_{q=1}^{N}a_{p}(y)a^*_{q}(y).
462: \end{equation}
463: Note that the sum of the arbitrary functions of $y$ on the right hand side of
464: eq.(36) becomes possible due to the structure of eqs. (1). In practice one
465: can choose ${\sum_{j=1}^{N}a_{j}(y)\over 2\sqrt{b(y)}}$ conveniently, for
466: example, as a combination of algebraic and hyperbolic functions. With N=2,
467: choosing the functions as
468: \begin{equation}
469: {a_{1}(y)\over \sqrt{b(y)}} = 2\mbox{sech}(m_{1}y+\delta_{1}),
470: {a_{2}(y)\over \sqrt{b(y)}} = 2\mbox{sech}(m_{2}y+\delta_{2}),
471: \end{equation}
472: where $m_{1}$, $m_{2}$, $\delta_{1}$ and $\delta_{2}$ are parameters, solving eq.(29),
473: one can obtain
474: \begin{eqnarray}
475: b(y)&=&\mbox{exp}\bigg(-\frac{1}{l_{1R}}((8l_{1R}^4-6\omega_{1R}^2)y
476: +8\kappa(\mbox{tanh}(m_{1}y+\delta_{1})\nonumber \\
477: &&+\mbox{tanh}(m_{2}y+\delta_{2})+2\int\mbox{sech}(m_{1}y
478: +\delta_{1})\mbox{sech}(m_{2}y+\delta_{2})dy))\bigg).
479: \end{eqnarray}
480: Then, the induced two dromion solution is given by
481: \begin{eqnarray}
482: \chi_{2}&=&e^{i\hat{\psi_{1I}}}(\mbox{sech}(m_{1}y+\delta_{1})+\mbox{sech}(m_{2}y+\delta_{2}))
483: \mbox{sech}\bigg[l_{1R}(x+2l_{1I}t) \nonumber \\
484: &&-\frac{1}{2l_{1R}}\bigg((8l_{1R}^4-6\omega_{1R}^2)y
485: +8\kappa(\mbox{tanh}(m_{1}y+\delta_{1})\nonumber \\
486: &&+\mbox{tanh}(m_{2}y+\delta_{2})+2\int\mbox{sech}(m_{1}y
487: +\delta_{1})\mbox{sech}(m_{2}y+\delta_{2})dy)\bigg)\bigg]
488: \end{eqnarray}
489: and is shown in Fig. 2. One can identify the mutual influence of
490: one dromion over the other from the additional terms occurring in the square
491: bracket of eq.~(40).
492: \begin{figure}[!ht]
493: \begin{center}
494: \includegraphics[width=0.75\linewidth]{fig2}
495: \end{center}
496: \caption{Snapshot of the induced two dromion solution of the Mel'nikov
497: equation (see Eq. (40))}
498: \end{figure}
499:
500:
501: If one chooses both the functions $a_{j}$, $j=1,2$, algebraically,
502: \begin{equation}
503: {a_{1}(y)\over \sqrt{b(y)}} = {2\over{(m_{1}y+y_{10})^2+1}},
504: {a_{2}(y)\over \sqrt{b(y)}} = {2\over{(m_{2}y+y_{20})^2+1}}
505: \end{equation}
506: or one of the functions to be algebraic and the other one to be hypebolic
507: \begin{equation}
508: {a_{1}(y)\over \sqrt{b(y)}} = 2\mbox{sech}(m_{1}y+\delta_{1}),
509: {a_{2}(y)\over \sqrt{b(y)}} = {2\over{(m_{2}y+y_{20})^2+1}}
510: \end{equation}
511: one can generate different kinds of induced lump-lump or dromion-lump solutions,
512: respectively.
513:
514: We have also tried to generate more general two soliton solutions by choosing
515: \begin{equation}
516: g^{(1)}=a_{1}(y)e^{\hat{\psi_{1}}}+a_{2}(y)e^{\hat{\psi_{2}}}
517: ,\hat{\psi_{1}}=l_{1}x+\omega_{1}t, \hat{\psi_{2}}=l_{2}x+\omega_{2}t,
518: \end{equation}
519: which lead to a condition $l_{1}=l_{2}$ thereby reducing to our original
520: form (36) for $N=2$. This is also true for $N>2$. Thus we believe that the
521: solution (36) constitutes the most general localized solution we could
522: construct for the Mel'nikov equation through our procedure.
523:
524: \section{Conclusion}
525: In this paper, we have pointed out the interesting fact that the (2+1)
526: dimensional Mel'nikov equation admits exponentially localized solutions
527: of different classes. We have also checked its integrability through Painlev\'e analysis.
528: We have in particular constructed localized dromion solutions and obtained
529: new classes of localized solutions such as dromions with spatially varying
530: amplitude and induced dromions.
531:
532: %\noindent {\bf Acknowledgement}
533: \section*{Acknowledgement}
534: The work of C.S. and M.L. form part of a Department of Science and Technology,
535: Govt. of India sponsored research project.
536: %\section*{References}
537:
538:
539: \begin{thebibliography}{11}
540: \bibitem{ref1}
541: Boiti M, Leon J J P, Martina L, Pempinelli F. Scattering of localized
542: solitons in the plane. Phys. Lett. A 1988; 132: 432-439.
543: \bibitem{ref2}
544: Fokas A S, Santini P M. Dromions and a boundary value
545: problem for the Davey-Stewartson I equation. Physica D 1990; 44: 99-130.
546: \bibitem{ref3}
547: Konopelchenko B G. Solitons in Multidimensions. Singapore: World Scientific;
548: 1993.
549: \bibitem{ref4}
550: Radha R, Lakshmanan M. The (2+1) dimensional sine-Gordon equation;
551: integrability and localized solution. J. Phys. A: Math. Gen. 1996; 29:
552: 1551-1562.
553: \bibitem{ref5}
554: Lou S Y. Generalized dromion solutions of the (2+1)-dimensional KdV equation.
555: J. Phys. A: Math. Gen. 1995; 28: 7227-7232.
556: \bibitem{ref6}
557: Tang X Y, Lou S Y, Zhang Y. Localized excitations in (2+1)-dimensional
558: systems. Phys. Rev. E. 2002; 66: 046601.
559: \bibitem{ref7}
560: Radha R, Vijayalakshmi S, Lakshmanan M. Explode-decay dromions in the
561: non-isospectral Davey-Stewartson (DSI) equation. J. Nonlinear Math. Phys. 1999;
562: 6: 120-126.
563: \bibitem{ref8}
564: Radha R, Lakshmanan M. A new class of induced localized coherent
565: structures in the (2+1) dimensional nonlinear Schr$\ddot{o}$dinger equation.
566: J. Phys. A: Math. Gen. 1997; 30: 3229-3233.
567: \bibitem{ref9}
568: Mel'nikov V K. On equations for wave interactions. Lett. Math. Phys. 1983;
569: 7: 129-136.
570: \bibitem{ref10}
571: Mel'nikov V K. Reflection of waves in nonlinear integrable systems.
572: J. Math. Phys. 1987; 28: 2603-2609.
573: \bibitem{ref11}
574: Conte R, Fordy A P, Pickering A. A perturbative Painlev\'e approach to
575: nonliner differential equations. Physica D 1993; 69: 33-58.
576: \bibitem{ref12}
577: Grammaticos B, Ramani A, Hietarinta J. A search for integrable bilinear
578: equations: The Painlev\'e approach. J. Math. Phys. 1990; 31: 2572-2578.
579: \bibitem{ref13}
580: Hase Y, Hirota R, Ohta Y, Satsuma J. Soliton Solutions of the Mel'nikov
581: Equations. J. Phy. Soc. Japan 1989; 58: 2713-2720.
582:
583:
584: \end{thebibliography}
585: %\endbib
586:
587:
588: \end{document}
589: